FMT2000 ctsm53 with 52 params (f.e30_alpha04a.FMT2000BgcCrop.52params.lmwg_dev_77)

[slevis-lmwg]

Description:
Same as #76 but with alternate params file, so as to come close to what used to be ctsm52 params.

---

Case directory:
Locally:
./create_newcase --compset 2000_CAM70%MT_CLM60%BGC-CROP_CICE%PRES_DOCN%DOM_MOSART_SGLC_SWAV_SESP --res ne30pg3_ne30pg3_mg17 --case ~slevis/cases_LMWG_dev/f.e30_alpha04a.FMT2000BgcCrop.52params.lmwg_dev_77 --run-unsupported

---

Sandbox:
Locally:
/glade/work/hannay/cesm_tags/cesm3_0_alpha04a

---

user_nl_ changes:
paramfile = '/glade/u/home/slevis/cases_LMWG_dev/f.e30_alpha04a.FMT2000BgcCrop.52params.lmwg_dev_77/ctsm5.3.012.Nfix_params.c241127.nc'

---

SourceMods:
None

---

Diagnostics:
https://webext.cgd.ucar.edu/FMT2000/f.e30_alpha04a.FMT2000BgcCrop.52params.lmwg_dev_77/lnd/f.e30_alpha04a.FMT2000BgcCrop.52params.lmwg_dev_77.11-20-f.e30_alpha04a.FMT2000BgcCrop.CTRL.lmwg_dev_76.11-20/setsIndex.html

---

Output:
Output (if still available):
Initially:
/glade/derecho/scratch/slevis/f.e30_alpha04a.FMT2000BgcCrop.CTRL.lmwg_dev_77/run
Later:
/glade/derecho/scratch/slevis/archive/f.e30_alpha04a.FMT2000BgcCrop.CTRL.lmwg_dev_77
Finally at:
/glade/campaign/cgd/tss/people/slevis/f_cases/f.e30_alpha04a.FMT2000BgcCrop.CTRL.lmwg_dev_77

---

Contacts:
@slevis-lmwg

---

Extra details:

[slevis-lmwg]

Both cases now running.

@wwieder also... I adjusted permissions to my $HOME. Please LMK if you can now see again directories and files you could previously.

[wwieder]

Yes, I can poke around your home directory now, Sam 😃

[slevis-lmwg]

@wwieder I started the diagnostics script for this case against the control:
/glade/u/home/slevis/lnd_diag/run/lnd_template4.2.39.casper.LMWGdev77-76.csh
and I have the next one ready to go right after:
/glade/u/home/slevis/lnd_diag/run/lnd_template4.2.39.casper.LMWGdev77-b116.csh

[slevis-lmwg]

@wwieder based on your post in #76 I will change the script that's waiting to
/glade/u/home/slevis/lnd_diag/run/lnd_template4.2.39.casper.LMWGdev76-b116.csh

[wwieder]

@linniahawkins can you briefly summarize what parameter differences are in this case compared to #76?

[wwieder]

Well we can juice up tropical LAI! 20 years in here's what LAI looks like in this case, with values > 12 in the Amazon!

And the difference from controls, which I'd argue still seem reasonable.

both plot are for Jan in year 23 of the simulations that started with identical initial conditions.

[slevis-lmwg]

Land diags finished overnight, and I just posted them.

[slevis-lmwg]

@adamrher
given the FMT2000 conversation in the CESM Proj. meeting, is there much benefit from using FMT2000 or would you recommend switching to F2000 for faster turnaround? Or it's difficult to know without seeing both :-)

[linniahawkins]

@linniahawkins can you briefly summarize what parameter differences are in this case compared to #76?

I can't access the case directory for #76. Can you point me to the parameter file used in that simulation? (mainly wondering whether the FUN_fracfixers bug was corrected yet or not).

[slevis-lmwg]

@linniahawkins I will see what I can do about access and will post an update in a few minutes (and sorry).

[slevis-lmwg]

@linniahawkins I think you should have appropriate permissions soon, and in the meantime case 76 used:
paramfile = '/glade/campaign/cesm/cesmdata/inputdata/lnd/clm2/paramdata/ctsm60_params.c240822.nc'
while case 77 as you know of course used:
paramfile = '/glade/u/home/slevis/cases_LMWG_dev/f.e30_alpha04a.FMT2000BgcCrop.52params.lmwg_dev_77/ctsm5.3.012.Nfix_params.c241127.nc'

[adamrher]

given the FMT2000 conversation in the CESM Proj. meeting, is there much benefit from using FMT2000 or would you recommend switching to F2000 for faster turnaround?

@slevis-lmwg you could increase throughput by a factor of 2 or so by running the "FLT" equivalent (L58). The tropospheric climate is ~identical to "FMT" (L93). I would not recommend F2000 (L32).

[linniahawkins]

Summary of parameter differences between #76 and #77 :
76:
ctsm5.3 LAI tuning: (reduce LAI bias, while maintaining ET in tropics)
deadveg tuning: FUN_fracfixers=1 for arctic shrubs and grasses (pft 11,12)
Fangs fire tuning: froot_leaf=0.5 for deciduous tropical trees and C4 grass (pft 6,14)

77:
undo deadveg tuning: FUN_fracfixers=0.25 for pft 11,12
undo fire tuning: froot_leaf=1.5 for pft 6,14
undo LAI tuning for arctic grasses and shrubs (pft 11,12): higher LAI, ET
undo LAI tuning for broadleaf evergreen tropical trees (pft 4): higher LAI, ET
increase C4 grass: relax nitrogen limitation by reducing costs.

The objective was to dramatically increase LAI and ET in the tropics to see if we could see a response in precipitation in Fcases (not to find a new parameter set for the coupled model).

The parameter changes in the arctic were made to adapt to the new Nfix temperature response curves, but I don't think that parameterization is in these runs so we also ended up increasing high latitude LAI.

[linniahawkins]

Even though we induced a large increase in tropical LAI, the change in amazon ET was modest (5-10 W/m2) which was expected. The net result of a small decrease in Precip (<1mm/day) over the amazon/congo is a little surprising.

I don't know what to make of the large decrease in precip over indonesia.

Bringing back C4 grasses looks to have had a bigger effect on ET (red blobs south of the Congo, and in eastern Brazil), slightly increasing precip.

[olyson]

Test post

[dlawrenncar]

Thanks @linniahawkins. Interesting that the parameter changes to the Arctic and tropical PFTs induced a significant looking change in mid-latitudes. See, for example, the impact on FSDS in JJA. Seems like a quick conversation tomorrow might be helpful, though I am not seeing a good time.

[slevis-lmwg]

@slevis-lmwg you could increase throughput by a factor of 2 or so by running the "FLT" equivalent (L58). The tropospheric climate is ~identical to "FMT" (L93). I would not recommend F2000 (L32).

@adamrher the FLT case defaults to
ncdata = '/glade/campaign/cesm/cesmdata/inputdata/atm/cam/inic/se/FLT_L58_ne30pg3_IC_c220623.nc'
Do you approve?

[adamrher]

Yep!